![Institutional Repository [SANDBOX]](images/dias.logo.png "Institutional Repository [SANDBOX]"){kind=logo}
Institutional Repository [SANDBOX]
Technical University of Crete
Technical University of Crete
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/70300711-F354-4B42-A1C9-8959A19DDB36 | ||
| Year | 2019 | ||
| Type of Item | Diploma Work | ||
| License |
|
||
| Bibliographic Citation | Argyris Moulios, "Real time stock forecasting", Diploma Work, School of Electrical and Computer Engineering, Technical University of Crete, Chania, Greece, 2019 https://doi.org/10.26233/heallink.tuc.81593 | ||
| Appears in Collections |
“Big Data analytics” is the method by which we look at big data to reveal hiddenpatterns, incomprehensible relationships, and other important data that can be used to solve decision-making problems. In recent years there has been an increasing interest in big data due to their rapid growth and since it covers different application domains.An important area of Big Data application is in the financial system and morespecifically in the field of stock prediction. In this case, the data streams can bemassive and continuous. Many times in stock forecasting, it is crucial the data to be processed and draw conclusions in real-time in order to end up with useful forecasts which can provide us with in time solutions in decision-making problems. The aim of this diploma thesis is to identify time-delayed correlated pairs among thousands of shares, in a real-time and distributed way, in order to extract information that will be used in their prognosis. More specifically, we are interested in extracting information about the course of shares that affects the course of others, the time lag and the correlation degree in this phenomenon. To achieve this goal, it is critical to ensure the scalability of the techniques we are about to use, in order to maintain a reasonable time for the results outputs, despite the increasing volume of incoming data.Therefore, we implemented the algorithm covering locality sensitive hashing (CLSH) without false negatives, on top of the distributed system Apache Storm, which we will analyze and present as well as the experimental results of this study.
Copyright © DIAS 2013
